Skip to content
v0.0.0-devApache-2.0

Your project's memory: the sessions that happened, the durable observations drawn from them, and the plans. Skills and tools for reading why the code is the way it is, and for recording what you learn.

myco

Myco is this project's memory — the sessions that happened, the durable observations (spores) drawn from them, and the plans. It holds why the code is the way it is, which the code itself cannot show. Use it before a design decision, when debugging something non-obvious, when a structure looks arbitrary, when picking up work someone else started, and after learning something worth keeping.

myco-handoff

Carry a session's working intention to another session or another agent. PREPARE compresses why we are here, what we are doing, the gotchas, the dead ends and the decisions already closed into a digest of at most 1500 tokens, saved on a Myco plan. RECEIVE takes that plan id, rehydrates the context, marks the referenced plans in progress, loads the suggested skills and resumes the work.

myco-okf

Create or maintain an OKF-conformant project wiki — a portable, git-committed markdown knowledge base that lives in the repository rather than in Myco. Turns what Myco holds, the spores and sessions and plans, into human-readable documentation that ships with the code. Any agent can run the procedure; it needs nothing beyond the ordinary Myco tools.

myco-outcome-tasks

Author and debug the tasks Myco's own agent runs. A task is one prompt with declared expected evidence, dispatched to a worker: no phases, no turn budget, no orchestrator. It declares the gate that admits it, the tools its run may call, its schedule if it has one, and how the server decides afterwards that it actually did the work. Getting those four declarations consistent is the whole job.

myco-rules

Keep AGENTS.md minimal, durable and canonical across every agent that reads it. Adds, trims and audits project rules without letting the rules file decay into a dump of whatever was true last week.

myco-server-debugging

Work out why a Myco deployment refused something, lost something, or stopped doing something. The server answers a caller's own fault as a terminal refusal carrying a named code, and its own faults as a retryable failure — telling those two apart is the first move in every investigation, because only one of them is worth retrying and only one of them is your bug.

myco-server-provisioning

Provision, update, roll back and destroy the Cloudflare Deployment — the Worker front door — with myco server <verb> --target cloudflare. One verb, from artifacts the binary carries: no source checkout on the machine that deploys, and no container runtime anywhere.

myco-settings

Change a Myco deployment's settings safely, and add a new setting correctly. Deployment settings are one tier with one writer: every change admits the leaf, authorizes the actor, checks the leaf's own rule, persists it with its actor, and re-arms whatever schedule it moved — in that order, in one module. Bypassing that path is how one of those five steps quietly goes missing.

myco-setup

Finish setting Myco up on this machine. The plugin alone gives you skills and the Myco tools; the binary adds session capture, plan capture, import and the worker. This skill checks what is present, installs the binary with the user's consent, signs in or redeems a join code, wires the agent's hooks and verifies the result — and names the outcome plainly when an install cannot proceed.